Scholarlink

<p>This study compares the use of Generative AI and a visual knowledge mapping tool (Mytelemap) as cognitive support tools in undergraduate programming education. In a counterbalanced within-subject design conducted over six weeks, students completed programming exercises across four C# topics while alternating between the two tools. Performance was evaluated using a rubric assessing scenario relevance, program correctness, code quality, reflection, and testing. The results indicated that students achieved higher exercise scores when using Generative AI for the programming tasks. However, students rated Mytelemap more positively for ease of use, visualization, and perceived accuracy of knowledge maps. A mismatch was therefore observed between perceived tool quality and measured task performance, suggesting that the tools may support different aspects of the learning process. These findings should be interpreted cautiously because the study evaluated short-term programming exercise performance rather than long-term learning outcomes. The intervention was conducted over a limited six-week period with a single cohort of 30 students, and the evaluation focused on rubric-based task scores rather than measures of retention or transfer. The results therefore provide an early comparative evaluation of how generative AI and visual knowledge mapping tools support programming task performance in classroom exercises, highlighting potential complementary roles for these tools in programming instruction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Learning to code with AI or maps? a comparative study of cognitive support tools in programming learning

  • 期刊论文

  • 发表日期：

    2026年7月1日

查看全文
  • Athitaya Nitchot,
  • Lester Gilbert

摘要

This study compares the use of Generative AI and a visual knowledge mapping tool (Mytelemap) as cognitive support tools in undergraduate programming education. In a counterbalanced within-subject design conducted over six weeks, students completed programming exercises across four C# topics while alternating between the two tools. Performance was evaluated using a rubric assessing scenario relevance, program correctness, code quality, reflection, and testing. The results indicated that students achieved higher exercise scores when using Generative AI for the programming tasks. However, students rated Mytelemap more positively for ease of use, visualization, and perceived accuracy of knowledge maps. A mismatch was therefore observed between perceived tool quality and measured task performance, suggesting that the tools may support different aspects of the learning process. These findings should be interpreted cautiously because the study evaluated short-term programming exercise performance rather than long-term learning outcomes. The intervention was conducted over a limited six-week period with a single cohort of 30 students, and the evaluation focused on rubric-based task scores rather than measures of retention or transfer. The results therefore provide an early comparative evaluation of how generative AI and visual knowledge mapping tools support programming task performance in classroom exercises, highlighting potential complementary roles for these tools in programming instruction.